Language development in typical children follows specific developmental sequences and demonstrates inherent biases at certain stages. It is not known to what degree language development in children with autism follows the same developmental rules or achieves language competence via different routes. The proposed study investigates two interrelated questions: First, is the process of language acquisition in autism similar to that of typically developing children? And second, what do language comprehension measures reveal about the process and products of language acquisition in children with autism? The proposed longitudinal study will use a language comprehension paradigm (Intermodal Preferential Looking, IPL) that has revealed important information about language in typical children 15 and 36 months of age but has not been used lo study autism. It combines the ability to display dynamic events with the necessity of only minimal response (visual fixation) and may reveal strengths and weaknesses not shown on standardized testing. Participants will be 18-36 months old, recently diagnosed with autism/PDD and about to begin a home Applied Behavior Analysis (ABA) program. A baseline assessment before treatment will be followed by assessments at 4-6 month intervals for 2 years, plus a final assessment at age 5-7 years. IPL will be used o examine multiple aspects of language such as word order in sentences and verb learning; the children will also be videotaped during naturalistic interaction with their mothers to provide information about spontaneous language production. Standardized assessments of the children's linguistic, nonverbal, and social functioning will also be obtained. Analyses will focus on (a) the degree to which comprehension exceeds production, (b) the degree to which preferential looking reveals more language ability than is demonstrated with traditional response methods (point, touch) and tests, (c) whether language development in ABA-created children with autism follows typical progressions and inherent biases, and (d) how early patterns of language development might predict ASD children's language abilities at ages 5-7. Results have the potential to provide useful new information to aid treatment by pinpointing subtle deficits and strengths in language, as well as to illuminate possibly different developmental pathways in language acquisition. [unreadable] [unreadable]